Data Attribution Reading Group
Methods and Applications in the Era of Generative AI
This is a Summer 2024 reading group on data attribution, which aims to measure the influence of individual training data points on machine learning models trained on them. The reading group will first cover several classic data attribution approaches, and then move on to focus on recent developments in the era of generative AI.
Schedule
| Meeting date | Topic | Presenter | Recording |
|---|---|---|---|
| 2024/06/22 | Influence Function | Jiaqi Ma | Link |
| 2024/06/29 | Counterfactual Subset Prediction | Jiaqi Ma | Link |
| 2024/07/06 | Data Shapley | Ziqi Liu | Link |
| 2024/07/13 | LoGra & Unrolled Differentiation | Juhan Bae (Guest) | N/A |
| 2024/07/20 | Data Attribution for LLMs (1) | Yijun Pan & Jin Huang | Link |
| 2024/07/27 | Data Attribution for LLMs (2) | Junwei Deng | Link |
| 2024/08/03 | Factual Knowledge Attribution in LLMs | Qirun Dai & Shixuan Liu | Link |
| 2024/08/10 | Data Attribution for GenAI Copyright | Ting-Wei Li & Junwei Deng | Link |
| 2024/08/17 | Data Attribution and for Image Generation | Jingyan Shen | Link |
| 2024/08/24 | Misc Data Attribution Methods | Yuzheng Hu | Link |
| 2024/08/31 | Data Attribution for Broader NLP | Yubo Zhang | Link |
| 2024/09/07 | Agnostic Data Attribution | Pingbang Hu | Link |